The major objective of this research is to study a new method of structured EEG analysis using computerized techniques. Capitalizing on large existing EEG data bases and on-going separately funded clinical research investigations, the work seeks a better understanding of the relationship between visual scoring of the EEG and syntactic EEG classification strategies. A second objective is to study the classification methodologies derived with data obtained from three collaterally-funded clinical research projects. The methods employed are derived from syntactic and multivariate pattern recognition strategies. The component elements of a current prototype EEG analysis system are being systematically evaluated and research conducted on methods for optimizing and discerning the characteristics of the analysis system. Computer results are studied and analysis methods modified using quantitative visual scoring of EEGs by three electroencephalographers. Ultimately, the results of these investigations should allow structured characterization of EEG features in the selected research areas studied and provide a sensitive and accurate classifier that will mimic visual scoring of EEG records. The research will contribute to a better fundamental understanding of methods of quantitative EEG analysis and, as a practical by-product, provide improved methods for analysis of EEG abnormalities in clinical medicine.